gardner the real tragedy of the commons

7/28/2019 Gardner the Real Tragedy of the Commons

http://slidepdf.com/reader/full/gardner-the-real-tragedy-of-the-commons 1/31

The Real Tragedy of the CommonsAuthor(s): Stephen M. GardinerSource: Philosophy and Public Affairs, Vol. 30, No. 4 (Autumn, 2001), pp. 387-416Published by: Blackwell PublishingStable URL: http://www.jstor.org/stable/3557968

Accessed: 17/02/2009 06:28

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available athttp://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless

you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you

may use content in the JSTOR archive only for your personal, non-commercial use.

Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at

http://www.jstor.org/action/showPublisher?publisherCode=black .

Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed

page of such transmission.

JSTOR is a not-for-profit organization founded in 1995 to build trusted digital archives for scholarship. We work with the

scholarly community to preserve their work and the materials they rely upon, and to build a common research platform that

promotes the discovery and use of these resources. For more information about JSTOR, please contact [email protected].

Blackwell Publishing is collaborating with JSTOR to digitize, preserve and extend access to Philosophy and

Public Affairs.

http://www.jstor.org

http://www.jstor.org/stable/3557968?origin=JSTOR-pdf

http://www.jstor.org/page/info/about/policies/terms.jsp

http://www.jstor.org/action/showPublisher?publisherCode=black

http://www.jstor.org/action/showPublisher?publisherCode=black

http://www.jstor.org/page/info/about/policies/terms.jsp

http://www.jstor.org/stable/3557968?origin=JSTOR-pdf



STEPHENM. GARDINER The Real Tragedyof the Commons

In two celebrated and widely anthologized articles, as well as severalbooks, the biologist Garrett Hardin claims (a) that the world population

problem has a certain structure: it is a tragedy of the commons; and, (b)

that, given this structure, the only tenable solutions involve either coer-

cion or immense human suffering.' In this article, I shall argue for two

claims. First, Hardin's arguments are deeply flawed.2 The population

Partsof this articlewerepresentedto the InternationalSocietyforEnvironmentalEthicsat the AmericanPhilosophicalAssociation (EasternDivision) meeting in Atlanta n 2001,

and to the AustralasianAssociation forProfessionalandAppliedEthicsconference in Bris-bane in 2000. I'd like to thank my audience on those occasions and, in particular,Dale

Jamieson,AndrewBrennan,and my APAcommentator,MargaretHolmgren.I have alsobenefited from discussions with TimBayne,Simon Blackburn,Roger Crisp,PhilipPettit,DavidSchmidtz and HenryShue;and from the written comments of an anonymous re-viewer for Philosophy& PublicAffairs.Some of the work on this article was completedwhile I was a visitor at HertfordCollege,OxfordUniversity,and CornellUniversity,as the

recipientof an ErskineFellowship. would liketo thank all three institutions fortheirsup-port.I owe most to NicholasSturgeon,who not only set me on this path,but also provideda fineexampleof how to do analyticphilosophywell.

1. "Tragedy of the Commons," Science 162 (1968): 1243-48; "Living in a Lifeboat," Bio-

science 24 (1974): 561-68. See also Living Within Limits (Cambridge: Cambridge University

Press,1993);The OstrichFactor:OurPopulationMyopia(Oxford:OxfordUniversityPress,1999).Some estimation of the influence of Hardin'sworkand the esteem with which it isheld in the scientific community can be found in JoannaBurgerand MichaelGochfield,"TheTragedyof the Commons30YearsLater,"Environment December1998): -27;the re-

sponses in "TheTragedyof the Commons Revisited,"Environment(March1999):4-5, 45;and Hardin's "Extensions of 'The Tragedy of the Commons'," Science 280 (1998): 682-83.

Burgerand Gochfieldinclude a graphof the citationindex from1968 o 1997 howing per-sistentlevels of over 100citationsper yearamongnaturaland social scientists.

2. Hardin'sworkhas, of course, been criticizedbefore, but I have not found the criti-cism Iwillmake in the literature.First,most of the criticism is empirical,especiallythat of-feredby proponents of a benign demographictransitionhypothesis, who maintain that

2002 byPrincetonUniversityPress.Philosophy&PublicAffairs30, no.4



Philosophy&PublicAffairs

problem as he conceives it does not have the structure of a commons;and even if it did, this would not necessitate the extreme responses he

canvasses. Second, nevertheless, much of Hardin'spessimism is justi-fied. Some environmental problems associated with population size do

have tragicstructures,althoughthese are of a differentformthan Hardin

envisions. Forexample, the problemof globalclimate change has an in-

tergenerationalaspect that makes it significantlyworse than Hardin's

commons, and for this reason (as opposed to Hardin's)extreme re-

sponses may be needed to avertenvironmentalcatastrophe.3

I. HARDIN'SANALYSIS

In 1804,after a wait of approximately wo millionyears, human popula-tion reachedone billion. One hundredtwenty-three yearslaterin 1927, t

topped two billion;thirty-threeyearslater,in 1960, three billion. By1974,

fourteenyears later,therewere fourbillion people; thirteenyears on, in

1987, ive billion; and twelve years afterthat, in 1999,six billion.4This is

an amazingrate of progression.BillMcKibbenreportsthat if the world's

populationhad increased

bythe same number each

year throughoutits

history as it did in 1994,then thinkingbackwardsfrom its currenttotal,the proverbialAdam and Evewould have to have started out in 1932.5

There are some positive signs. Forexample, the rate of increase in the

number of humans appears to be slowing down.6Nevertheless, since

this rate is being applied to an expanding base of people, the absolute

number of birthswill only come down to what it is today by the second

quarterof this century.Furthermore,because people are living longer,

populationwill be curbedbydevelopment.Unfortunately, aluablethoughit is, this work

does not directlyaddressHardin'smain argument,which is not empiricalbut theoretical.Second, much of the theoretical criticism of Hardin revolves around his metaphors in

"LifeboatEthics,"not those in "Tragedy f the Commons."For a good discussion of the

majorviews and theirweaknesses,see JesperRyberg,"Populationand ThirdWorldAssis-

tance," Journal ofApplied Philosophy 14 (1997): 207-19.

3. The climatechange problem s a useful illustration ortwo reasons:becauseit seems

to be the most seriousenvironmentalproblemwe face at present;and because somethinglike the same incentive structureappears o some degree n other environmentalproblems.

4. United Nations, Departmentof Economic and SocialAffairs,PopulationDivision,

1998Update http://www.census.gov//ipc//www/worldpop.html).5. BillMcKibben,"Reachinghe Limit,"New YorkReviewofBooks(May29,1997).6. The U.S.Bureauof Censusestimates the peak to have been in 1962and 1963at 2.19

percent per year.

388



TheReal Tragedyof the Commons

the total population will still be rising at the midcentury mark, and willthen likely be around nine billion; that is, by around 2054, global popu-lation will be fifty percent larger than it is today. Hence, the problem of

population growth is very much with us.

Population is a problem because the increased absolute number of

people, and the rate of increase itself, are both likely to have a severe im-

pact on the planet. Extra people place extra demands on food, water,and energy supply, and their activities cause environmental damage. So,it is important to understand what or who7 is causing the problem, and

perhaps thereby determine what if anything can be done about it.Hardin offers some dramatic answers to these questions. First,

he claims that the population problem has a special structure-it is a

commons problem-and that this structure "remorselessly generates

tragedy."8The tragedy is that, left to their own devices, people have largefamilies, causing misery to themselves and their communities and un-

told damage to the environment. Second, Hardin sees the problem as

one primarily caused by, and affecting, those in the developing nations.

Third, he argues that the only available solutions are severe. In one ar-

ticle, heargues

that we should abandon the United Nation's declara-

tion that freedom to reproduce is a fundamental human right.9 In-

stead, Hardin thinks, we should use coercive instruments to prevent

people from reproducing, or reproducing more than is wanted. In an-

other article, Hardin argues that the affluent nations should refuse to

assist their poorer neighbors in times of humanitarian crisis. Instead,

he endorses Tertullian's claim that we would be wise to think of "pestil-

lence, famine, wars, and earthquakes" as "prun[ing] away [their] luxu-

riant growth."'lHardin's idea is that the earth provides a corrective to the problem of

population through natural catastrophe. But, he thinks, human interfer-

7. This is not to say that the issue is one of whom to blame, in the moralsense. Hardinthinksblamingmisses the point. In "Lifeboat,"e says, "theconcepts of blame and pun-ishment are irrelevant.The question is, what are the operational consequences of estab-

lishing aworldfood bank?" p.563).InLivingWithinLimits, his is on his assumptionthat"eachhuman being, like every other animal, is genetically programmedto seek its own

good,"so that" t]hetragedy s broughton not byindividual sin ('greed'),butbythe systemitself" (p. 218).

8. "Tragedy,"p. 1244;emphasis added.

9. Ibid., p. 1246;Ostrich, p. 145.

o1. "Lifeboat,". 564.

389



Philosophy & Public Affairs

ence has disrupted the natural mechanism. Misguided altruism, in theform of the welfare state and food aid to overpopulated countries, has

meant that the costs of overpopulationno longer fall on those who have

the children.These institutions have createdwhat Hardincalls 'atragedyof the commons'.

Thetragedy s illustratedwith an example.Hardinasks us to imagine a

groupof farmersgrazingcattle on common land. He says:

As a rational being, each herdsman seeks to maximize his gain. Ex-

plicitly or implicitly,more or less consciously, he asks, "What s the

utilityto me of adding one more animal to my herd?"This utilityhas

one negative and one positive component.

(1) The positive component is a function of the increment of one

animal. Since the herdsman receivesall the proceeds from the sale

of the additionalanimal, the positive utility is nearly+1.

(2) The negative component is a function of the additional over-

grazingcreated by one more animal. Since, however,the effects of

overgrazingare sharedby all the herdsmen, the negative utilityfor

anyparticulardecision-makingherdsman is

onlya fraction of-i.

Addingtogether the component partialutilities, the rational herds-

man concludes that the only sensible course for him to pursue is to

add another animal to his herd. And another.... But this is the con-

clusion reached by each and everyrational herdsman sharinga com-

mons. Therein is the tragedy.Each man is locked into a system that

compels him to increase his herd without limit-in a world that is

limited. Ruin is the destination towardwhich all men rush, each pur-

suing his own best interest in a society that believes in the freedom of

the commons. Freedomin a commons bringsruin to all."The commons, as Hardinunderstands it, seems to have two features.

First, t is a multipersonprisoner'sdilemma.12 econd, it governsa singlecommon resource.Inthis article,I shallfocus on the firstfeature.13

11. "Tragedy,"p. 1244.Repeated in LivingWithin Limits, pp. 217-18.

12. See Nick Griffin, "Lifeboat USA, Part 1,"International Journal of Moral and Social

Studies 3 (1988): 230-31.

13. Thesecond featurealsoposes difficulties or Hardin.Environmentalproblemsasso-

ciated with population growthcan impactresourcesat various levels (local, national,re-

gional and global).Presumably,hen, the best candidate for a single common resource s

the biosphereas a whole. ButHardindoes not take this approach.Instead,in "Lifeboat,"

390



TheReal Tragedy of the Commons

II. THE PRISONER'SILEMMA

A prisoner'sdilemma is a situationwith a certain structure.In the stan-

dardexample,two prisonersare about to stand trialfor a crime they are

accused of committing together.'4Each faces the following proposition.He can either confess or not confess. Ifboth confess, then each gets five

years.Ifneitherconfesses, then each gets one yearon a lesser charge.But

if one confesses and the otherdoes not, then the confessor goes free,and

the nonconfessor gets ten years.Neither knows for sure what the other

will do;but each knows that the other faces the same choice situation.Given this scenario,each personhas the followingpreference ranking:

(1) I confess, the otherprisonerdoesn't. (Go free)

(2) Neither of us confess. (1year)

(3) Both of us confess. (5years)(4) I don'tconfess, but the otherprisonerdoes. (lo years)

This situation is usuallyexpressedwith a diagram:

Person B

Person

A

don't confess confess

don't confess i, 1 (2nd, 2nd) 10, o (4th, 1st)

confess o, 10 (1st, 4th) 5, 5 (3rd, 3rd)

The reason why the situation is called a dilemma is as follows. Sup-

pose I am one of the prisoners.I cannot guaranteewhat the other pris-oner will do, and have no effectivemeans to make it that I can do so. So I

he rejectsa "spaceshipearth"metaphorin favorof speakingof the richcountries as indi-vidual lifeboats that should defend themselves againstthe lifeboats of the poor countries.

But it is not clearwhat the ecologicalgroundscould be fortakingindividualcountries as

single common resources.For urthercriticisms,see WilliamW.Murdochand AllenOaten,

"Population nd Food:Metaphorsandthe Reality,"Bioscience25 (1975);eprinted n Louis

Pojman,EnvironmentalEthics,2ded. (Belmont,Calif.:Wadsworth,1998),357-61.14. The title and illustrationareattributed o AlbertTucker,who used them to popular-

ize ideas developed by MerrillFlood and Melvin Dresher in investigatingglobal nuclear

strategy.Formore information,see Stephen Kuhn,"The Prisoner'sDilemma,"StanfordEncyclopedia of Philosophy, http://setis.library.usyd.edu.au/stanford/entries/prisoner-

dilemma/#Bib'

391



Philosophy &Public Affairs

need to consider each possibility. Suppose he confesses. Then it is betterfor me to confess also (since five years in jail is better than ten). Supposehe does not confess. Then it is better for me to confess (since going free

is better than one year in jail). So, whatever he does, I should confess.

But the situation is exactly the same for him. So, reasoning in the same

way I do, he will confess also. Hence, the outcome will be that both of us

confess (getting five years each). But this yields a suboptimal outcome.

Each of us prefers the outcome that comes from us both not confessing(one year each) over the outcome that comes from us both confessing

(five years each).15For our purposes, the problem can be (roughly) characterized as

follows:

PDi: It is collectively rational to cooperate: each agent prefers the out-

come produced by everyone cooperating over the outcome produced

by no one cooperating.

PD2: It is individually rational not to cooperate: when each individual

has the power to decide whether or not she will cooperate, each per-son (rationally) prefers not to cooperate, whatever the others do.

PD1and PD2 generate the paradox as follows. In prisoner's dilemma sit-

uations, each individual has the power to decide whether or not she will

cooperate. Hence, given PD2, if each person is individually rational, no

one cooperates. But this means that each person ends up with an out-

come that she prefers less than another outcome that is available. For,

according to PD1, each prefers the cooperative over the noncooperativeoutcome.'6

15. But neither of us can get there as things stand.Supposeone of us thinks,It'sa pris-oner'sdilemma, so we should not confess. Then, the other person knows that we knowthis. Butifhe thinks we arenot going to confess, the rationalthingfor him to do is to con-

fess, since thisgiveshim a betteroutcome. Remember hatthe previousreasoningshowedthata prisonershould confess,no matterwhat the otherprisonerdoes.

16. The prisoner'sdilemma structureis interestingfor both theoretical and practicalreasons. The theoretical reason is that it involves a paradoxof rationality. t shows that insome situations, individualsreasoning purely on the basis of self-interest can be led tomake decisions that are suboptimalin terms of self-interest.(Strictly-speaking,he prob-lem does not depend on self-interestedmotivationper se, but might arise for any value

systemwith a similarstructure,ncludingsome moralviews. See DerekParfit,ReasonsandPersons[Oxford:OxfordUniversityPress,1984],pp. 55-56, 95-110.)Thepracticalreasonis

that thereare real worldsituations that have this structure.

392



The Real Tragedy of the Commons

Hardin's description of this kind of situation as a tragedy is apt. Forwhat happens is more than simply a bad thing. The situation as it stands

drives people by an inexorable process towards a situation that is worse

by their own lights, and away from a situation that is better. Indeed, it is

the very same values that make cooperation preferable that drive each

agent away from it. In the classic example, prisoners want less jail time,

that is why they collectively prefer cooperation; but their desire for less

jail time also makes it individually rational for them not to cooperate.

III. COERCIONAND POPULATION

Now, suppose the population problem were a prisoner's dilemma. What

would follow? Hardin suggests two things. In "Tragedy of the Com-

mons," he says that we should mutually agree on the use of coercive in-

struments to prevent people from reproducing, or reproducing more

than is wanted. In "Lifeboat Ethics," he argues that the affluent nations

should refuse to assist their poorer neighbors in times of humanitarian

crisis.

These recommendations are difficult toaccept.7

Indeed, there is a

strong moral presumption against Hardin's solutions. For they implythat the rich countries should deliberately allow hundreds of millions of

people to die, when the rich could help the poor at relatively small cost

to themselves. This conflicts with even a very weak principle of benefi-

cence.18 It may also be resisted on other moral grounds.'9

17. Rybergalso arguesthat even if Hardin s rightabout the severityof the problem, it

may not necessitate his solutions. Ryberg'sargument relies on the idea that aid to the

"overpopulated"ountriescoupledwithregulationof populationvia famine and environ-mental disaster may still be best on Hardin'sconsequentialist grounds. See Ryberg,pp.

212-15.Hence, he disputes Hardin's mpiricalclaims about the overall results of the pris-oner'sdilemma. By contrast,I take issue with Hardin'sassumptions about what followsfrom the dilemma situation itself.

18. Aprincipleof beneficence maybe conceived of as eithera principleof moralgood-ness (orvirtue), or as a moralrequirement.HereI directlyintend the former,as benefi-cence is less controversialas a moral ideal (say,of charity),than as a moral requirement.But I also believe that the weakprincipleI have in mind is plausibleas a requirement.

19. Forexample,some mightarguethat, in withholdingaid, the rich countries seem totreatthe people who starve as a resultdisrespectfully.For the starvingseem to be treatedin waysto which they could not in principleconsent, and so can be conceived of as beingcoerced.(Onemight also saythattheyaretherebyused bythe rich as a means to solving a

widerproblem.)Hence, theiractions conflict with majortenets of Kantian(and arguably

commonsense) morality,which claim we should not treat others in ways to which we

393




Given the presumptions against Hardin's solutions, surely we shouldthink on moral grounds that if almost anything else will work, then that

is what we should do. This makes it relevant to consider alternative waysin which prisoner's dilemma situations might be resolved, ways that do

not involve coercion. At least three are pertinent here. The first is to

change people's motives, and so alter the preference structure that gen-erates the problem. Hence, one might try to make people value some as-

pect of the situation in a new way. For example, fairly recently, many

people have been made to disparage styrofoam cups to the extent that

they have such a strong preference to drink from a reusable cup, thatthey are willing to carry one around with them. Or one might try to make

people value cooperation itself.

The second solution is to appeal to broad considerations of self-interest.

For example, it is well known that the dynamics of a prisoner's dilemma

can be changed if the parties will meet again in other bargaining situa-

tions in the future. Ifwe know that we must make a bargain again, we are

much more likely to give up some gains from noncooperation now in ex-

change for the expectation of gains to be made from an overall strategyof

cooperation.The third solution is to appeal to a sense of fair play, and in particularto the notion of reciprocity. This is present in almost all societies, and

supports a social attitude of rebuke to those who do not cooperate: theyare socially shunned. This too can work to solve some problems without

the need for coercive state interference.20

Now the problem for Hardin is that, even if the population problemwere a prisoner's dilemma, it seems plausible to think that any one of

could not in principle consent, and that we should never treat them merelyas a means.

Theseclaimsare,of course, controversial, ven amongstKantians.Butan approachto aidproblemswhich makes moves of this sort has been defended by OnoraO'Neill,"EndingWorldHunger" n Tom Regan,ed., Mattersof Life and Death, 3d ed. (Belmont, Calif.:

Wadsworth,1993),and Facesof Hunger:An Essayon Poverty, ustice,and Development(London: Allen &Unwin, 1985).

20. In the originalstandard titfortat' solutions to the prisoner'sdilemma,the secondand thirdstrategies it together.RobertAxelrodreported hat "thetwo requisitesforcoop-erationto thrive are that the cooperationbe based on reciprocity,and that the shadow ofthe future s important enough to make this reciprocitystable." Axelrod,TheEvolutionofCooperation NewYork:BasicBooks,1984]; election reprinted n Ethics,PeterSinger,ed.

[Oxford:Oxford University Press, 1994], pp. 88-92 at p. go.) For egoists, these conditions

hold when there is prolonged interactionovertime and where this interaction holds the

promiseof greatbenefits to both sides.

394



TheReal Tragedy f the Commons

these solutions (or some combination) might help to resolve it. The ap-peal to broad self-interest looks especially promising. The economic

costs of having children are huge; and the noneconomic costs are also

frequently high. Though, arguably, he noneconomic benefits of parent-hood aresimilarly arge,it seems more than possible to persuade peopleeither to forgothese altogether (byhaving no children of their own, and

makingdo with being biologicalornonbiological aunts oruncles, forex-

ample), or to have fewer of these benefits (byhaving fewer children), in

exchange for the savings incurredby not having children.21ndeed, ar-

guably,this is at least part of what has already happened in the devel-oped countries,and has reduced the number of children there.

However,Hardinseems to treat the possibility of such solutions with

outrightdisdain.Inparticular,he seems to thinkthat an appeal to broad

self-interest will not work because self-interest is far too deeply tied to

the production of children for any change in reproductivebehavior to

occur.22 nd he regardsan appeal to fairness as not only ineffective, but

also self-eliminating.These attitudes of Hardin'ssuggest that he regards the population

problemnot

justas a

prisoner'sdilemma, but a

prisoner'sdilemma that

depends on circumstances of a particularlydeep and intractable kind.

So, in order to understand Hardin'sattitudes, we must look at his argu-ments for the claim that population is a prisoner'sdilemma.

IV. POPULATIONSAPRISONER'SILEMMA

If population is a prisoner'sdilemma, then it must have the followingfeatures:

Population-PDi:It is

collectivelyrational to have smaller families:

Each agent prefers the outcome produced by everyone havingsmallerfamilies overthe outcome produced by everyone having largefamilies.

21. Foran overview,see AndrewHacker,"The CaseAgainstKids,"New YorkReviewofBooks, vol. 47, no. 19 (November 30, 2000): 12-18.

22. Hardinsays,"Ruin s the destination towardwhich all men rush,each pursuinghisown best interest in a society that believes in the freedomof the commons,"("Tragedy,".

1244, mphasisadded);about the populationproblemin particularhe says,"the ndividualbenefits as an individualfromhis abilityto deny the truth [aboutthe population tragedy]

even though societyas a whole,of which he is a part,suffers" "Tragedy,". 1244).

395




Population-PD2: It is individually rational to have a large family:when each individual has the power to decide whether or not she will

have a large or small family, each person (rationally) prefers to have a

large family, no matter what everyone else does.23

The first problem for Hardin, which is widely noted, is that, on the as-

sumption that most actual people are individually rational to this ex-

tent, Population-PD2 seems to be empirically false. This is true both of

its content and its form.

With regard to content, the raw data is remarkably consistent. First, as

is often pointed out, there has been a significant transition in Western

Europe over the past loo years or so. All agree that Western Europe is be-

low replacement level, and that this is almost true now in North Amer-

ica.24There was a spurt of population growth with industrialization, but

now this is over. Furthermore, this happened without coercion. So, Pop-ulation-PD2 is not true of the behavior of people in these nations.25 Sec-

ond, the UN reports that global fertility rates have fallen significantly in

the recent past (to 2.7 children per family in 1998, from 5 in the early

1950s); and, furthermore, this reflects a decline in fertility in all regions

of the world (in the last twenty-five years, 6.6 to 5.1in Africa, 5.1to 2.6 inAsia, 5.0 to 2.7 in Latin America and the Caribbean).26So, Population-PD2 seems unlikely to be true of the behavior of the world as a whole ei-

23. Hardin ends to speakin terms of "overbreeding,"nd seems to mean bythis 'hav-

ing as many healthy, survivingchildren as possible'.But his general argument requiresonlythe more modest 'have a size of familyabovereplacement evel',and this is what I in-tend here. (Ihaveavoidedthe use of'breeding'as this termmayseem offensive,andin anycase encouragesthe animalistic connotations Hardin tries to give human behavior,con-notationsI resistbelow.)

24. The NorthAmericanpopulation figuresare complicated by large immigrationto

both the United Statesand Canada.Withoutimmigration,North Americais likelyto beslightlybelowreplacement.

25. It is not trueeven in those nations officially n favorof morerather han fewer chil-dren.Forexample: Italyis officiallyCatholic;but nineteenth-century Italyhad a fertilityrate of five childrenper woman;but now it is down to significantlybelow that necessaryforreplacement.Indeed,the United Nationspredictsverysubstantial mmigrationwill be

necessaryto maintainits currentpopulation.26. UnitedNations,1998Update.Thisdata s used by optimistsaboutpopulation n sup-

portof their view that therewillbe a benign demographic ransition o lowertotalpopula-tion. The pessimists may arguethat these declines arecoming too late (andthey may be

right).Butthis does not help Hardin's rgument, or this is an empiricalquestion,to be an-sweredbythe data.ButHardindoesn'tprovideanydata. Heprovidesa conceptualargument

to show that the populationproblem s aprisoner'sdilemma,and thatcoercion srequired.

396




ther: all seem to be followingthe downward trend toward smallerfamilysize.27

As well as being in stark conflict with the raw data, Population-PD2also takesan overlysimplisticform. Itreduces the question of reproduc-tion to the issue of familysize. But even in the developing nations, mat-

ters are much more complex than this. For example, the recent Indian

census has suggested that ultrasoundtechnology is enabling Indians to

follow the Chinese in abortingfemale fetuses at abnormally high rates.

InIndia,then, it seems clearthatsex plays a significantrole in reproduc-

tive choice, and one that affects family size. And this fact is clearlyrele-vant forpopulation policy.28

The second problem for Hardinis that Population-PD2 seems inde-

pendently implausible. Although Population-PDi seems plausible

enough (itis undoubtablytruethat everyone is worse off if all have largefamilies),what makes it in my interests to have a large family whatever

anyone else does? Consider the following:For a prisoner'sdilemma, we

need this incentive structure:

CoupleB CoupleBhasmerelyreplaces a large family

Couple A 2nd, 2nd 4th, ist

merely replacesMaintaincurrentpopulation

CoupleAhas a 1st,4th 3rd,3rd

large family

Populationexplosion

27. For a good discussion of the data, see ElizabethWillot, "PopulationTrends," nDavid Schmidtz and ElizabethWillot, eds., Environmental Ethics:WhatReallyMatters,WhatReallyWorksOxford:OxfordUniversityPress,2001).

28. Amajorcause of the problemseems to be (a)that these societies have no universalsocial security systems, and (b) that women are effectivelyrelievedof all responsibilitiesfor theirown familywhen they marry.This tends to make the rationalityof havinga child

dependon its sex, andperhapsthe rationalityof havingan additional child depend on thesexof the precedingchildren.See CeliaW.Dugger,"ModernAsia'sAnomaly:The GirlsWhoDon'tGetBorn,"TheNew YorkTimes,Week n Review,Sec. 4 (May6, 2001): 4. The issue isdiscussed in depth byAmartyaSen,"MissingWomen,"BritishMedicalJournal304 (March

7, 1992): 587; and Development as Freedom, (NewYork: Anchor, 1999), chap. 8.

397




But why would we think that couples have these preferences, so that the

preferred outcome for all is to have a large family, and what is feared

most of all is for others to have large families while we do not? Under

moderately favorable conditions, it seems more plausible that some

people will prefer replacement, or close to replacement, no matter what

everyone else does; that very few people will prefer having lots of chil-

dren just for the sake of it; and that some will prefer no children (or one).

Indeed, the empirical evidence from the developed countries suggeststhat overall, taking everyone's preferences into account, uncoerced deci-

sions produce a level of reproduction that is below the replacement levelfor a whole society.

To resist this argument, Hardin has to claim that it is always to the ad-

vantage of the individual to have a large family.29But why should we be-

lieve this? It is not clear. But Hardin seems tempted by the view that it is

because it is biologically advantageous to have a large family. For exam-

ple, he says:

If each human family were dependent only on its own resources; ifthe children of improvident parents starved to death; if thus, over-

breeding brought its own "punishment" to the germ line-then therewould be no public interest in controlling the breeding of families.

But our society is deeply committed to the welfare state, and hence is

confronted with another aspect of the tragedy of the commons. In a

welfare state, how shall we deal with the family, the religion, the race,

or the class (or indeed any distinguishable and cohesive group) that

adopts overbreeding as a policy to secure its own aggrandizement? To

couple the concept of freedom to breed with the belief that everyoneborn has an equal right to the commons is to lock the world into a

tragic course of action.30

29. Hardinseems to realize this. He says that the independentherdsman "darenot re-frain"romoverloading he commons,because if he did so he would "suffermore" han a"selfish"ne who does overgraze"Lifeboat,". 562); urthermore,f he did refrain romover-

grazing, heherdsmanwouldbe (correctly) ondemnedas a "simpleton""Tragedy,". 1246).30. "Tragedy,".1246; mphasis n original.As is oftennoted, thisseems to be aninaccu-

ratedescriptionof the motivationof those in above-replacement ountries.Peoplein somesocieties have reason to have lots of childrenbecause thereis no one else to lookafter hemwhentheybecome old and there s high infantmortality.This makesit riskynot to havelotsof children,fromthe point of view of self-interest.This is actuallyworse than a prisoner'sdilemma situation. Forhere the parentsarelikelynot to preferthe constrainedoutcome,

becausetheyfearabandonment n old agemore thangeneralpopulationproblems.

398




Now, this argument faces two problems. First, it is concerned withgroups and germlines. Itsays nothing about the interests of the individ-

uals involved.Therefore, t does nothing to justifythe claim that it is al-

ways to the advantage of the individual to have a large family.Second,this claim would be plausible only if one posited a strongcorrelationbe-

tween the biological interestsof a germline and the self-interest of indi-

viduals carryingthat germline.But,on any plausible theory of the inter-

ests or well-being of an individual,an individual'sself-interest does not

consist in, nor is it dominated by, even his or her own biological inter-

ests, let alone the biologicalinterests of the group.31Butperhaps Hardin has a second argument that does not depend on

people'sinterests. When arguing against the idea that an appeal to con-

science might solve the population problem (my 'fairplay suggestion'),he says:

Peoplevary.Confrontedwith appeals to limited breeding, some peo-

ple will undoubtedly respond to the plea more than others. Those

who have more childrenwillproduce a largerfraction of the next gen-eration than those with more susceptible consciences. The difference

will be accentuated, generation by generation.... The argument as-sumes that conscience or the desire to have children (no matter

which) is hereditary-but hereditaryonly in the most general formal

sense. The result will be the same whether the attitude is transmitted

through germ cells, or exosomatically.... The argument has been

stated in the context of the population problem, but it applies equallywell to any instance in which society appeals to an individualexploit-

ing a commons to restrainhimself for the general good-by means of

his conscience. To make such an appeal is to set up a selective system

that workstowardthe elimination of conscience fromthe race.32Hence, Hardinargues that reproductiverestraintwill be eliminated bynaturalselection: those who practice restraint will have fewer descen-

dants than those who do not, and since the attitude of restraint s trans-

31. There s also a questionabout whether one can makesense of biologicalinterestsin

evolutionary erms (andalso,perhaps,whetherone can make sense of the notion of bio-

logicalinterests).Such an approach s triedby GaryVarner, nNature's nterests?:nterests,AnimalRights,and EnvironmentalEthics Oxford:OxfordUniversityPress,1998).Butnotethat even Varnerbelieves thatindividualshave psychologicalinterestsin additionto their

biologicalinterests,and that the psychological ntereststrumpthe biological.

32. "Tragedy,". 1246.

399




mitted between generations,this attitudewill become progressively esscommon.

This argument faces serious practical, empirical, and theoretical

problems. The practical problem is that natural selection works over

very long time-scales. Hence, it is unlikelyto happen fast enough to pre-vent the appeal to conscience workingfor a while. And perhaps a while

is all we need worryabout, if the benign demographic thesis is correct.

The empirical problem is that if Hardin were right, we would expect

people alreadyto have the desire foras many children as possible, since

we would expect the selection procedureto have been atwork forgener-ations.33 ut the empiricalevidence suggests the opposite:global fertilityrates arefalling.The theoreticalproblemis the assumption that the atti-

tude of restraint s transmitted between generations.Thereis simplyno

reason to believe that peoples' consciences will be the same as theirpar-ents',nor in particularthat they will have the same attitude toward re-

production.Indeed, the empiricalevidence stands squarelyagainst it.

So, I conclude that Hardin'sanalysis of the population problem as a

prisoner'sdilemma is untenable. It relies on flawed assumptions about

human motivation that rest onextremely

dubiousappeals

to evolution-

ary biology.

V. TOTAL NVIRONMENTALMPACT

The discussion so farhas shown that Hardin'sanalysisof the population

problem as a tragedy of the commons is fatally flawed, and even if it

were not, Hardin's olutions would not be justified. But there is one re-

spect in which Hardin'saccount is close to the truth.Atleast one aspectof the population problem does have a tragicstructure;and extreme so-

lutions maybe requiredto overcomeit.

33. Thatis, we would expect people already o have the desire for as many childrenas

wouldmaximize the chance of the genes being passedon. Againstthis,Hardinwouldpre-

sumably arguethat the historical naturalpruninghad an effect on reproductivemotiva-

tion, but one which is being underminedby the more recent welfare state and foreignaid

programs.But here it is worth pointing out (a) that those countries without the welfare

state have sustainedhigh birth ratesratherthanheld them back,and (b)that the declinein birth rates has occurred in rich countries even given the introduction of welfare.

(Against b],Hardinmight arguethatit is too earlyto tell what will happen in these coun-

tries,giventhat welfare s fairlyrecent;but stillthe empiricalevidence does not lookprom-

isingforhim.)

400




Tobegin, it is worth observingthat the importantissue about popula-tion is not how many people there are.In itself, this tells us nothing and

threatens little. The real issue is the environmental impact that peoplehave, inparticularon the so-called carryingcapacityof the earth.34 hus,if there is an issue of population, it is the issue of how the number of

people in the world interactswith the environmentalimpact per person,to produce the total environmentalimpact of humanity.

On this issue, however,there is reason to believe that the most impor-tant variable is the environmental impact per person, not the total hu-

man population.35 f this is right, the benign democraphic thesis turnsout not to be so encouraging after all. For that thesis suggests that the

price of a decrease in the absolute number of people is development.But development as we presentlyunderstand it requiresadditional en-

ergy consumption, and this in turn tends to involve an increase in the

environmental impact per person.36 For the major sources of energy

currently supporting the developed countries, especially oil and coal,come with significant environmental impacts.) Indeed, the develop-ment necessary for the less developed nations to reach a stable or even

declining populationwould, on

present technologies,involve a cata-

strophic increase in energy consumption, and so in the environmental

impactper person.37

34. Determiningthe carrying apacity s a difficultandextremelyvalue-ladenbusiness,which raises philosophicalissues of its own. Fora helpful discussion, see JoelE. Cohen,HowManyPeopleCantheEarthSupport? NewYork:Norton,1995),chap. 12, and app. 3and4.

35. See Griffin, "Lifeboat,"pp. 223-4.

36. Globalenergy consumption rose from 21 to 318 exajoulesbetween 1900 and 1988

(SeeWilliam C. Clark, n ManagingPlanet Earth (NewYork:Freeman, 1990),p. 87. BillMcKibbenreportsthatpercapita consumption had risenfrom less than 1megawatt-hour

per person in 1800to 19megawatthoursperperson in 1989(McKibben,TheEndof Nature[New York:Anchor, 1989], p. 2).

37. Forexample,in 1997, ndia and China had commercialenergyuse per capita of 479and 907 kg of oil equivalent respectively,compared to 3, 863 and 8,079 for the United

Kingdomand United Statesrespectively. SeeWorldBank,WorldDevelopmentIndicators

2000.)

Hardinrecognizes that energy consumption per capita is a useful partialmeasure of

qualityof life. He also points out the sharpdifference between figuresfor developed andless developedcountries. Buthe refused to see this as aglobal problem.Instead,he arguesfor a localizing principle:that parochialdistribution of resources should be matched byparochial consumption;and so sees the problem as a local one. (Forexample,he all but

suggests that Bangladeshshould reduce its population from 104million to 2.7 million,

rather than bemoan its lack of resources.)But there are two problems for Hardin here.

401




This news is bad enough, but there is worse to come. For the problemof an escalating environmental impact per person seems to have a

structure similar to the one that so worries Hardin in his misdiagnosis of

the problem of world population. Indeed, the structure of this problemis in some respects worse than the one Hardin envisions, and so more of

a tragedy.

VI. THEINTERGENERATIONALSSUE

There are both inter- and intragenerational problems involving environ-mental impact. Consider first the intergenerational problem. Much of

the damage inflicted by energy consumption is by current generationson future generations.38 For example, global climate change is caused

mainly by emissions of carbon dioxide, but the effects of these emis-

sions on the global temperature suffer from an important time-lag. The

lifetime of carbon dioxide in the upper atmosphere is over 100 years, so

that the full (cumulative) effects of current emissions will not be felt un-

First,his defense of the principle ("theremust be some sort of fragmentationof adminis-trative asks")s too weak eitherto justifythe principleitself,or the conclusionshe is pre-paredto draw from it. Second,Hardinadmitsthat the principledoes not applyto airand

water,where a globalapproach s needed. But,as I shall stressbelow,the most serious en-vironmentalproblemscausedby currentenergyconsumptioninvolveatmosphericpollu-tion, and so areglobal. (See Hardin"CulturalCarryingCapacity:a BiologicalApproach oHumanProblems,"Bioscience36 (1986): 99-606,esp. 602-5.)

38. Futuregenerationsare often definedas those futurepeople whom those presentlyalivewill not live to meet (see, for example,Avnerde-Shalit,WhyPosterityMatters:Envi-ronmentalPoliciesand Future GenerationsLondon:Routledge,1995],p. 138).Forthe sakeboth of simplicity,andof makingclearthe fundamentaldynamics,I follow thatusagehere.

However, futuregenerations' s also often used to referto those not presentlyalive,and it

is importantto emphasize that many of the points I will make apply in a more nuancedway to this less restrictednotion. Thisis because the crucial ssue is not when those whowill live when we have gone appear,but the extent of our present concern fortheirwell-

being. (This s whatgivesa point evento the morerestrictiveuse of the term "future ener-ations.") ndeed,it is worthnotingthat much of what is said about futurepeople in the re-stricted sense appliesnot only to futurepeople in the less restrictedsense, but also (in a

graduatedway) to people alreadypresentlyalive.Foreven when there is overlap,and wecareabout the well-beingof at least some of the people who remain afterwe aredead,thatconcern tends to be less than our concern for people around now (evenwhen the same

people are at issue) andto decline overtemporaldistance. (Such ssues areimportant o afull account of the dynamicsof the globalwarmingproblem,since there one wouldwantto saymore aboutthe medium-rangeeffects of climatechange.Buthere I amofferingonly

a basicoutline,forillustrativepurposes.)

402



The Real Tragedyof the Commons

til the beginning of the twenty-second century. By contrast, the benefitof burning fossil fuels, the energy thereby produced, is consumed by the

present generation. Hence, whereas the present generation both causes

the environmental damage and reaps the rewards, most of the costs fall

on future generations.39This suggests that the current generation has a

powerful self-interested reason to carry on polluting, and the future

generations a powerful self-interested reason for wanting that pollutionto stop.40

Now, it is true that this problem has a tragic structure, and is a com-

mons. But it is not a tragedy of the commons in Hardin's sense. It isworse. Present and future generations both have an interest in the

earth's relative climatic stability, and therefore in its capacity to absorb

the by-products of fossil fuel emissions. But, of course, the future gener-ations have no control over what the present generation does with that

capacity; whether it stays within, uses up, or exceeds it. Indeed, theycould not in principle have any such control. They are not around to

present a claim, nor to represent their interests. Worse, even if they were

around, or even if someone could represent them, they would have no

bargaining power.There is

nothingthat

theycan offer the

present gener-ation in exchange for cooperation (or nothing that the present genera-

39. Twoqualificationsare worthmakinghere.First,not all of the rewardsaccrue to the

present generation.Some are passed on in the form of technological advances and in-creasesin the capitalstock. This raises the prospectthat futuregenerationsmightbe com-

pensated forthe damage they inheritthroughhavingbetter resourceswith which to dealwith them. Suchargumentsare no doubtwarranted n some cases:forexample,some de-

velopingcountries areprobablyrightto think that they do best to improvetheir economicinfrastructure ather han abate emissions at the moment, especiallysince the planet is al-

readycommitted to some warming.However, n general,the point is limited by such fac-tors as (a)that much of the benefitof emissions is not passedon but simplyconsumed; (b)

thattechnologyand capitalare far fromperfectsubstitutesforenvironmentalquality;and(c)thatthe precisephysicaleffectsof globalwarmingare ikelyto be unpredictable,severe,and possibly catastrophic(so that effectivedeployment of the inheritedbenefits to miti-

gatethem maybe extremelydifficult).Second, if the immediate effects of pollution are bad enough, the costs placed on the

present generation may,of course,be high enough to give them a self-interestedreasonnot to pollute.Butthispresentlyseems unlikely n the case ofCO2 missions. (Thisis espe-ciallyso because,even if the short-term uture mpactof climatechangeis significant,thiswill be causedby past emissions, and so is warmingto which the presentgenerationis al-

readycommittedthroughthe actionof past generations.)40. I leave aside the philosophicalproblem dentifiedby DerekParfitas the Non-Identity

Problem.SeeParfit,ReasonsandPersons, hap.16.Butnote than even Parfit hinksthatwe

can carryon talking n termsof harmsto futuregenerations(Ibid.,p. 367).

403



Philosophy Public Affairs

tion could not take anyway).So, what happens is completely up to the

present generation.And it has powerfulself-interested incentives to ex-

ceed the capacityand therebyalter the climate.

Given this analysis,other things being equal, it is reasonable to expectthat the commons will be deeply harmed by the present generation.Furthermore, he same reasoningwill applyto each futuregenerationas

it comes into being. Howevermuch it may deplorethe effects of the pre-vious generationsdecisions on it, each generationwill face the same de-

cision situation with respect to generations later than it. Hence, each

generation will pollute, and the pollution will continue as long as theearthcan bear it.

This analysis shares with Hardin's ragedyof the commons the claim

that a suboptimal outcome might eventuate even if collectivelyall gen-erationswould agreethat it would be better if the atmospherewere not

so exploited.Forthis agreementwould be based on what is better for the

human race as a whole, or better for each generationbar the firstif all

others do the same. But this qualification about the first generation is

extremely important. It makes the currentproblem worse than a pris-oner'sdilemma, for

restrictingpollutionis not betterforthe first

genera-tion. And then, providedeach generationtakes its startingposition as a

given, it is not better for subsequent generations either.So,we get a se-

quentialmotive towardsnoncooperation and overpollution.Inshort,the intergenerationproblemhas the followinggeneralfeatures:

Intergenerational-l: t is collectivelyrational for most generations to

cooperate: (almost) every generation prefers the outcome produced

by everyone restricting pollution over the outcome produced by

everyone overpolluting.

Intergenerational-2: t is individually rational for all generations not

to cooperate:when each generationhas the powerto decide whether

or not it will overpollute,each generation (rationally)prefersto over-

pollute,whateverthe others do.

This structureis worse than the prisoner'sdilemma structurein the

followingways. First,Intergenerational-i s worse that PDi because it is

not true that everyone prefers cooperation (hence, the "almost").The

firstgeneration capable of overexploitingthe atmosphere will preferto

do so, since it experiences no gains fromcooperation. Second, Intergen-

404




erational-2: is worse than PD2because the reason for it is deeper.In theprisoner'sdilemma, PD2 arises because there is infact no way for each

prisoneror herdsman to ensure that the behavior of others is beneficial

so long as his is beneficial. But here, Intergenerational-2arises because

the situation is such that it is in principle impossible for one generationto ensure that the behaviorof the others is beneficial so long as theirs is

beneficial.

The intergenerationalproblem not only has a more tragic structure

than Hardin's ragedyof the commons, it is also more resistant to solu-

tion. Earlier I argued, contrary to Hardin, that his problem could beresolved by standard noncoercive means, e.g., by appeal to broad self-

interest or to a sense of fairness. But here such solutions are problem-atic. First, the usual appeals to broad self-interest rely on there being

repeatedinteractions between the partieswhere mutuallybeneficial be-

havior is possible. But between present and future generations there is

neitherrepeatedinteraction(bydefinition, there is no interaction at all),nor mutual benefit (there is little future generations to can do benefit

present generations).4'Second, similardifficulties arise for standardap-

pealsto

reciprocity:uture

peoplecannot

reciprocate.Still,

perhapsone

could arguethat if the firstgeneration capable of serious long-term pol-lution unilaterallyrestricts its activities, then subsequent generationscan owe the obligation to their forefathers to restrict their pollution for

the sake of future generations. (Subsequent generations get a benefit

from not inheriting an overpollutedplanet, but then must, out of fair-

ness, pass this on, so that there is a kind of indirectreciprocity.)Unfortu-

nately, this solution faces both theoretical and practicalproblems. The

theoreticalproblemis thatwe need to assume that the initialgenerationmakes a pure sacrifice of self-interest,with no compensation. So, their

action cannot be justified by an appeal to (even indirect) reciprocity.Thepracticalproblemis that the antecedent is not satisfied:significantlong-term pollution has alreadyoccurred.42

These problems reflect the difference in structure of the cases alreadymentioned. For,in the prisoner'sdilemma case, most of the proposed

41. They mightcontinue certain traditionsand projects.See JohnO'Neill,"FutureGen-erations: Present Harms," Philosophy 68 (1993):35-51.

42. Perhaps t could be weaker.Perhapsall that is required s that some generationca-

pable of makinga differencemake the sacrifice.Still, he theoreticalproblemremains. Pre-

sumablyit is to be resolvedat least in partby appealto purelyaltruisticmoral reasons.

405




solutions rely on rearranging the situation so as to provide some kind of

guarantee of the behavior of others when one cooperates. But in the in-

tergenerational problem, there is no way to rearrange the situation in

this way.43 t is essential to the problem that the parties cannot interact,

and so that the future generations are in no position to benefit nor to en-

gage in reciprocal acts with their forbears.

VII. Two INTERPRETATIONSFTHEINTRAGENERATIONALROBLEM

Suppose for the moment that the intergenerational problem can beovercome, and people in the present generation agree that they have a

reason not to overpollute. Still there is a problem. For there remains a

collective action problem at the intragenerational level. Again, let us

take climate change as our example.

Now, there is some controversy about how to characterize the intra-

generational aspect of the global warming problem. In particular, there

seems to be some case for describing it as a Prisoner's Dilemma, and

some for describing it as a Battle of the Sexes.44A full account of these

43. Perhaps the point can be made clearerby looking at the preference structureswhich underliethe tragicsituation.Theintergenerationalproblembeginswitha prisoner'sdilemma structure:

1stpreference: pollute,previousgenerationdoes not.2ndpreference:NeitherInorprevious generationpollutes.3rdpreference: pollute, previous generationpollutes.4thpreference: don'tpollute,previousgenerationpollutes.

But for the firstgenerationcapableof seriousoverpollution, his becomes simply:

1stpreference: pollute.2ndpreference: don'tpollute.

So,thisfixes the thirdoptionforthe nextgeneration,and so on forsubsequent generations.44. Climatechangeis explicitlydescribed as a prisoner'sdilemmaby MarvinS.Sooros,

TheEndangeredAtmosphere: reserving GlobalCommons(Columbia,S.C.:Universityof

South Carolina Press, 1997);pp. 260-261. It seems to be implicit in many other analyses. Bycontrast,JeremyWaldronmakes the case for a Battle of the Sexes in an unpublishedre-

searchpaper kindly supplied to me by an anonymous referee for this journal. (Waldron,"Who s to StopPolluting?DifferentKindsof Free-RiderProblem," une-July1990.A Con-

tribution o a projectEthicalGuidelines orGlobalBargainsundertakenbythe Program nEthicsand PublicLife,CornellUniversitywith support from the RockefellerFoundation.

Waldron s applyinga generalapproachto collective action problems put forthby Jean

Hampton,"Free-Rider roblemsin the Production of CollectiveGoods,"Economicsand

Philosophy3 [1987]: 45-73;and Hobbes and the Social ContractTradition[Cambridge:

Cambridge UniversityPress, 1986].)The Battle of the Sexes analysis is also brieflysug-

406




matters would be beyond the scope of this article.45However,for currentpurposes we can make do with something more limited. I shall arguethat, even without decisively resolving the theoretical dispute, in prac-tice the intragenerational problem is best treated as if it were a pris-oner'sdilemma, and so as addinga furthertragicstructureto the generalproblemof globalwarming.46

To make this argument, I shall present a brief case for three claims:

first, that, seen from a long-term perspective, there is a strong case for

characterizingthe intragenerationalglobal warming problem as a pris-

oner'sdilemma; second, that if there is uncertaintyabout whether it is abattle of the sexes orprisoner'sdilemma, it should be treated as if it werea prisoner'sdilemma;and, third,that in any case the importance of clas-

sifyingit as a battleof the sexes instead of aprisoner'sdilemma is under-cut by the background presence of the intergenerational problem.

Let us begin with a preliminarycase foraprisoner'sdilemma interpre-tation. Consider the circumstances facing individual members of the

present generation (or individualfirms).47On the surface, these appearto generatea prisoner'sdilemma. Tobegin with, even if the members of

the present generation care about futuregenerations,

themarginal

ef-

fects of any individual'spollution are small. Furthermore, here arepow-erful incentives to cheat. First,not polluting involves an absolute sacri-

fice of energy consumption, and so of self-interest.Second, it is cheaperto pollute than not to pollute;and the cheaper energy is a comparative

gestedbysome remarksof NickMabey,StephenHall,ClaireSmith and SujataGupta,Argu-ment in the Greenhouse:The nternationalEconomicsof ControllingGlobalWarming Lon-don: Routledge, 1997), pp. 356-59; 409-10; and, for the specific issue of ratification of the

KyotoProtocol,ScottBarrett,"PoliticalEconomyof the KyotoProtocol,"OxfordReviewof

Economic Policy 14(1998):20-39; 36-37.45. I hope to pursuethe issues in more detail, and as applied to the KyotoProtocol in

particular,n a futurepapertentativelyentitled "TheGlobalWarmingTragedy."ome dis-cussion of the Kyotocontext is to be found in Barrett,"PoliticalEconomy."

46. There are two basic reasonswhy the structure of the intragenerationalproblem isunclear.First,solving the problemrequires resolvinga number of other collective action

problemsthat are subordinate,or otherwise closely related,to it. Second, because of the

uncertaintysurrounding he magnitudeand distribution of the variouscosts and benefitsof climatechange,the dynamicsof the intragenerational nd relatedproblemsare all sub-

ject to considerableuncertainty.47. Thewritersnoted above areallconcernedwithcountries.But Ibeginwith individu-

als for the sake of simplicity,since I believe that a prisoner'sdilemma analysiswould be

widelyacceptedat that level. I make a case forextendingthe analysisto countriesbelow.

407




advantage when one is producing things in the market place.48Third,

cheating might be difficult to identify. Carbon dioxide emissions, for ex-

ample, are very difficult to monitor and detect.49Fourth, policing will be

difficult in another way: so long as everyone cheats, cheating does not

harm anyone around at the moment; so there is no one presentlyaround with a powerful interest in stopping it. Indeed, as we have al-

ready seen, all the present people have a powerful interest in polluting,no matter what others do, and so have an incentive to turn a blind eye to

the pollution of others so long as others turn a blind eye to them.

If this is right, even without considering the intergenerational prob-lem, members of the present generation seem to have the following

preference structure:

1st preference: I pollute, you don't.

2nd preference: No one pollutes.

3rd preference: Everyone pollutes.

4th preference: You pollute, I don't.

But this, of course, is a prisoner's dilemma preference structure. Hence,

otherthings being equal,

left to their own devices each individual will

choose a strategy of polluting, since it gets better outcomes than not

polluting. But then everyone will pollute, which gives everyone onlytheir third preference, and not their second.

Now, it is tempting to suppose that the prisoner's dilemma analysis

applies not only to individuals and individual firms, but also to coun-

tries. But this extension may be questioned.50 For it may be said that it

48. This is in part because externality effects imply that compliance costs are likely to

be high, and in partbecause thereis an entrenchedcompetitive advantagebased on cur-rentpatternsof energyuse.

49. Things might be more promising if one considers the sources of carbon dioxide,namelythe fossil fuels themselves. Formost individualshave to buytheirfuel fromothers;and most countries have to import much of their fuel from elsewhere. (But note that:

[a]some countries do have significantfossil fuels of their own, e.g., China's oal reserves;[b]monitoringand enforcing imits on oil production,althoughone of the easier fuels to

monitor,has proveddifficult n the past, e.g., forOPEC,and for those imposing sanctionson Iraq;and [c]the emissions produced by any givenunit of, say,oil varyconsiderablyde-

pendingon use, in wayswhich makecalculationsof overallemissions levels difficult,evenwhen the agentsinvolvedarecompliant.)

50. Note that it would be possible to hold that countries confronta battle of the sexes,even if individualsconfront aprisoner'sdilemma.Therearetwo reasons.First, he numberof actorsis very different,makingit more plausible in the individual case that decisions

made byone partydo not directlyaffect decisions made bythe others.Second,on the one

408



TheRealTragedy f the Commons

neglects the fact that perhaps the collective good at stake can beachieved without everyonecooperating,and thereforethat it may be ra-

tional for a subgroupto produce the good alone.5'In this case, there will

be a situation where, forN players:

Partial Cooperation:Thereis a number M (such that M < N) that is

the minimum number of players whose cooperation is necessary if

some situation,which is bad for all, is to be avoided.

Marginal Cooperation: f the number of others who arewilling to co-

operate is just short of M, then a given party prefers to cooperate,since he is better off enjoyingthe benefits of cooperation and payinghis shareof the cost than he is in the situation where not enough peo-

ple cooperate.52

hand, a many-person collective action problem often has the character of a prisoner'sdilemma as far as a chooser is concernedif:(a)it involves an incrementalgood with indis-cerniblecontributions;or (b)a step good, and eitherthe threshold is indiscernibleor it isnot particularlyproximate n terms of my choice (i.e., if I do not perceive the crossingofthe thresholdas a close thing, then whathappens is that the effect of my contribution be-comes less and less discernible).But,on the otherhand, it has the characterof abattleof

the sexes for the individual if: (c) the good is incremental and contributions are dis-cernible;or (d)there is a discerniblyclose threshold.Arguably,he intragenerationalprob-lem facing individuals seems best characterizedby either (a) or (b): (b) with remotenessseems to be the waymost people see it;or else (a)with fatalistic attitude.Hence, they aremorelikelyto face a prisoner'sdilemma. But one might arguethat (c) or (d) is more likelyforcountries, since they are in a betterposition to gatherinformation,makeagreements,andguaranteecertain evels of reduction.Hence,countriesmightface a battle of the sexes.

(For hismethod of classification,IrelyonWaldron,who relies on Hampton.)51. Hampton says that many actual problems are mislabeled as prisoner'sdilemmas

when they reallyhave this structure.SeeHampton, "Free-Riding."52. Waldroncharacterizesthe global warmingproblem as follows in "WhoIs to Stop

Polluting,"pp.34-35:

(a)There are more than two players.Callthe number of playersN.

(b)There is a numberM (suchthatM < N),which is the minimum numberof playerswhose cooperationis necessaryif some situation,which is bad forall, is to be avoided.

(c) IfM playerscooperate,allN playersbenefit.

(d) Eachplayer prefersthe situation in which M players cooperatebut he is not one ofthem to the situationin which he is one of the Mcooperators.Inotherwords,coopera-tion is costly.

(e)Thegreaterhe numberofcooperators,hesmaller hecostto eachofcooperating.Thus

each, f he is acooperator, refershatthenumberofcooperatorsbe aslargeaspossible.

(f) Ifany cooperate,and the number of cooperatorsis less than M, then the coopera-

409




This situation is not a prisoner's dilemma. It is distinctive of a prisoner'sdilemma that each party does at least as well by not cooperating as he

does by cooperating, no matter what the others do. But in this situation,

if the number of other cooperators is just short of M, then a given playerdoes better by cooperating than not cooperating. This kind of situation

is usually termed a many-person battle of the sexes.53

Interpreting the intragenerational global warming problem as a battle

of the sexes has some initial appeal.54 Principally, it reflects the un-

doubted theoretical fact that it is total global emissions that need to be

tors suffer the cost and enjoy none of the benefits of cooperation.Each would prefernot to cooperatethan to be in this situation.

(g)If the number of others who arewillingto cooperateis just short of M,then a givenpartyprefersto cooperate, since he is better off enjoyingthe benefits of cooperationandpayinghis shareof the cost thanhe is in the situation where not enough people co-

operate.

So,I have pickedoutWaldron'sb)and (g).53. See Hampton,1987.The basic structureof the Battle of the Sexes can be gleaned

from the (unfortunatelysexist) story fromwhich it gets its name. A couple are decidingwhether to go out on a date. The man would most preferto go to see a game of baseball;thewomanwould most prefera tripto the theater.Butboth prefer o go on the dateratherthan to go to their own favorite event (or stay at home) alone. This yields the followingpreferencestructure:

Battleof the Sexes

Woman

Baseball Theater

Man

Baseball 1, 2 3, 3

Theater 4, 4 2, 1

Inthissituation,thereis no naturalequilibriumposition,but the two cooperative olutionsarebetterthan eithernoncooperativesolution.Hence,both playersare motivatedto com-

promisebecausebothwant to avoid thenoncooperativeoutcome.(Note: he phrase many-personbattleof the sexes'is unfortunatelyambiguousbetween a situation in which allpre-ferto cooperate,and one where allprefer hatsome cooperate.Hampton ntends thelatter.)

54. Forexample, t preservesthe insightthatclimatechangeinvolves some form ofcol-lective action probleminvolvinga strongincentive to tryto free ride (in this case, by get-ting others to form the subgroup).Furthermore,t appearsto be supported by some as-

pects of recentrealworldevents.Even n the face of the decisionbythe United States-theworld'ssingle largestemitter-to refuse to supportthe Kyotoagreement,the othercoun-tries of the world have made an agreementthat involves significantreductions in pro-

jected emissions for the other large emitters (prominently,the EuropeanUnion andJapan).Thissuggestsboth the Partialand MinimalCooperationclaims.

410




cut, so that in principle there is no physical reason why the costs of any

given percentage reduction need to be borne by all.5 Furthermore, if the

intragenerational problem were a battle of the sexes, this would be wel-

come news, practically speaking. For,unlike a prisoner's dilemma, a bat-

tle of the sexes may have internal solutions: that is, solutions that do not

require a change in the payoffs available to the players, or to the basic

psychology of self-interest.56 So, solving the intragenerational problemwould be easier than if it were a prisoner's dilemma.57

VIII. THE INTRAGENERATIONALROBLEMNPRACTICE

Nevertheless, there are some reasons for concern about the battle of the

sexes analysis. First, the idea of a fixed global percentage reduction does

not adequately capture the dynamic nature of the problem and the ex-

ponential nature of growth.58To begin with, any actual talk of stabiliza-

tion or a percentage reduction is relative to a fixed point (under Kyoto,

1990)where not all countries are making maximum use of fossil fuels (or,

more broadly, of CO2 emissions themselves), and it is quite likely that

none are.59So whatreally

needs to be looked at is thepercentage

reduc-

55. This is mentionedbyboth Mabeyet al. andWaldron,and seems to supportthe Par-tialCooperationclaim.

56. In particular,a battle of the sexes can be resolved if (a) a particularoutcome issalient to both parties as the likely cooperativeresolution (and so can act as a rallyingpoint), and (b) for the partywho must give up his most favoredoutcome to secure thesalientone, the expected utilityof concedingto the salient outcome is greater han that of

holdingout and so riskinga noncooperativeoutcome.

57. Again,this initiallyappearsto be supported byrecent events.

58. The basic battle of the sexes analysis gnoressome (further)mportantstrategic ac-torsthatemergeonce onlysome countriescomply, namelythe competitiveadvantages n-

ternal to not makingcuts. (Forexample,one's energy costs are cheaper than one's com-petitors'not only because they arespending money cuttingemissions that will be passedon to consumersin higher prices,but also because these effortswill lower the price of fos-sil fuels on the world marketsby reducingdemand for them and encourage emissions-intensive industriesto migrateto countries with no restrictions.)And it is possible that

they makeat least some of the attemptto reducetotal emissions self-defeating.(Interest-

ingly, Mabeyet al. consider the currentscenariounstable, for they believe that defectionthe United Stateswill makeenergy-intensive ndustries eave the cooperatingcountries insuch numbers that the benefits of cooperationfor those countries would be outweighedby their costs. See Mabey et al., Arguing, pp. 266, 299, 410.)

59. Of course, CO, is not the only greenhouse gas whose emissions are useful to hu-

mans; it is merely the one presentlymost important. (Methaneis also very important;

bringingthe slightlycomical problemof bovine flatulence to global prominence.) So, inthe end, an even morecomprehensive agreement s needed, as Kyotoenvisages.

411



Philosophy &Public Affairs

tion needed over the global total for maximum usage. This is unknown,as we don't know what possible uses of CO2emissions technological de-

velopments might make open to us. However, if, as a rough (and proba-

bly extremely conservative) estimate, we take current U.S. emissions as

the maximum per capita, based on current technology, we would get an

astronomical total from which to cut back to, say, 1990.60 Furthermore,

this means that the costs must be borne by anyone with the potential

ability to bring the global total above the optimal level of emissions. But

(depending on the figures above, and especially if the gains in terms of

self-interest from fossil fuel use have not yet been fully realized) thismeans everyone, or almost everyone.6' Without full cooperation one is at

best delaying global warming, not arresting it.62

If all this is right, it casts doubt on the partial compliance assumption.For it suggests one of two possibilities. If the size of the smallest player is

large enough that it alone could ensure that the ceiling is broken regard-less of what the others do, then the Partial Compliance Assumption is

false. Alternatively, if M is large, and close to N, then the Partial Compli-ance Assumption is misleading in the context. For the nations excluded

must have smallpopulations

and/or current emissions. Therefore,they

60. Veryroughly, or a world populationof 6 billion people, this would amount to 123

billion tons of carbon dioxide annually(based on 20.5 tons per capita emissions for theUnited States, the figure for 1995). This rises to 184.5 billion tons for a world population of 9

billion,which is projectedbythe U.S. CensusBureau ormid-century.Totalprojected emissions from energyconsumption for 1991was a mere 26.4 billion

tons. (Hence,continuation at 1991evels would allow a per capita averagefor emissions of

4.4 tons,which is only20 percentof the U.S.figure.)Hence,Chinaalone (withapopulationof 1.2 billion in 1995, rising to 1.5billion in 2025) could break the 1991 otal, even if (as would

neverhappen) other countriesemittednothing. (Slightly)morerealistically, nd simplify-ing considerably,even if all other countries cut back 20 percent from their levels in the

earlynineties-a huge reduction overprojected"business-as-usual" rowth-(to a total of

21.1 billiontons), a bloc of countries withverylow currentemissions and a combinedpop-ulation of over 250 million (i.e., 4 percent of the world'spopulation) could conceivablybreak the 1991 otal if they emitted at currentU.S. levels.As, of course,would the UnitedStates alone if it were the defector.(Figures rom theWorldResourcesInstitute,as citedbythe UNEPClimateChangeKit,http://www.unfccc.int/resource/iuckit/index.html.)

61. Thisis especiallyso if the marginalcosts of reductionget higherandhigher,which

theydo, so that the costs to a cooperating group goes up the more other countriespollute.This effect would be exacerbatedby a reductionin the price of fossil fuels broughtabout

bythe withdrawalof demandbythe coregroup.62. Andone might not be delayingit verymuch. Some say that the currentagreement

betweenEuropeandJapanwill only slowthe growth n atmosphericcarbon dioxidebysix

years. (This,of course, ignoresthe fact thatthe currentagreement only coversemissions

from 2008 to 2012, and envisions new targets for future years.)

412




will not be major players in the climate change problem, nor indeed inthe wider international arena.63The focus on partial compliance is thus

at best misleading.

Second, the Marginal Cooperation claim is also dubious. Consider it

first in relation to Hardin's commons. In Hardin's commons, it is not true

that only a subgroup of cooperators is required to secure the collective

good. Instead, any attempt by a subgroup to secure the optimal produc-

tivity of the commons will be undermined by the remaining noncooper-ators. For it is true both (a) that it is possible for any noncooperating

herdsman to increase his herd without limit, and (b) that each has an in-centive to do so, since the benefits accrue only to him whereas the costs

are shared by all. Indeed, this is just the fundamental dynamic that cre-

ates the commons problem in the first place.64The lesson here is that one has a prisoner's dilemma, not a battle of

the sexes, if it is possible for noncooperators to undermine the collec-

tive good produced by a subgroup, and if they have a strong enough in-

centive to do so. But just these features seem to be present in the global

warming case. For, given what we have already said, it is empirically

likelyboth that

noncooperatorscan

disruptthe

ceilingof emissions,

and that they have the incentive to do so.65Hence, the subgroup cannot

63. There are still problemsin leavingsuch nations out of an agreementoverthe longterm.Iftheycould increasepopulationbyreproductionorby immigrationovertime, then

they mightbecome a threatto the agreement;and there would be an incentive for individ-uals to emigratethereso long as an advantageexisted fordoing so:i.e., forCO2 ichwaysoflife. Thiswould implythatpartialcooperationis false after all.

Waldron'se) also suggeststhatN-Mwill be small;(e) states that the greater he numberof cooperators, he smallerthe cost to each of cooperating.Thus each prefersthat,if he is a

cooperator, he number of cooperatorsbe as largeas possible.Hence,each cooperatorhasan incentive to use anyother means at his disposalto put pressureon othercountries out-

side the subgroupto join. Since the cooperationof the largeand potentiallyeconomicallypowerfulnations is required or success (giventhe dynamic problem),it will be small and

economicallynot-so-powerfulnations who fall into N-M,but these are the most suscepti-ble to externalpressure rom these nations.Hence,it is to be expectedthatMwillbe equal,orveryclose, to N.

64. It is also worth noting that the more successful the subgroupis to begin with, the

greater ncentives to the noncooperatorsto do this, as the better the qualityof the com-mons theycan exploit.Furthermore,o maintain the commons, the M areunderprogres-sive pressureto do more and more, so that the costs on them in a dynamic context in-creasedramatically.

65. The theoreticaldifficultyhere may be in treatingtotal abatement as a step, ratherthanincremental,good. Ifabatement at some percentageof global1990 evels reallyrepre-

sented a threshold that benefited all, then one might argue that even noncooperators

413




enforce a ceiling, and would incur large costs if it tried. It seems, then,that the intragenerational global warming problem will be a prisoner'sdilemma.

Now, the above constitutes a strong case for a prisoner's dilemma in-

terpretation. Nevertheless, in the end, it is an empirical matter whether

this interpretation is correct, or whether global warming turns out to be

some other kind of collective action problem; and the empirical issues

are complex and shrouded in uncertainty. So, it is helpful to have a

secondary practical argument at hand. This goes as follows. If it is an

empirical matter whether global warming is, say, a battle of the sexes ora prisoner's dilemma, then it is at least possible that the problem is a

prisoner's dilemma (or worse). But then there are grounds to treat it as ifit were a prisoner's dilemma. There are three reasons.

First, if it is possible that global warming is a prisoner's dilemma, this

undermines the practical advantage of classifying the problem as a bat-

tle of the sexes, which was that such problems may have internal solu-

tions.66For if the problem may turn out to be a prisoner's dilemma, we

should not rely on internal solutions because they are unavailable for

prisoner'sdilemma. Instead, we should seek external solutions, such as

payoff changes and motivation changes, because these can resolve both

prisoner's dilemma and battle of the sexes. In general, if there is doubt

about whether a situation is prisoner's dilemma or battle of the sexes,

and if the noncooperative outcomes are potentially catastrophic, it

seems to make sense to employ those solutions that apply to both; if one

tries only battle of the sexes methods, then one might fail.

Second, even if the intragenerational problem were a battle of the

sexes, the relevance of this fact would be undermined by the back-

ground presence of the intergenerational problem. So far,we have sim-

would have an incentive not to disruptit. But this argumentwould be a mistake,for tworeasons.First, he influence of emissions is, as far asanyoneknows,or at least as faras it istreated n the models of the problem, ncremental,not a matter of threshold.Hence,thereis a threatat the margin.Second,and even more importantly, ince "disruption"ere sim-

plyinvolvescontinuingto polluteon a business-as-usualbasis (i.e.,to the extent thatit is a

benefit irrespectiveof concerns about climate),this "solution"would actually urnnonco-

operators nto cooperators.Forbearingsome of the costs of abatement heresimplymeansnot pollutingon a business-as-usual basis.

66. Notice the 'may'here. Internalsolutions require he identification of a salientsolu-tion. But in the case of global warming,this is underminedby the complex natureof the

problem.Inparticular, he identification of a relevantsubgroupis underminedby uncer-taintyaboutthe distributionof costs and bythe rangeof potentialcontributors.

414




ply assumed that problem away in this Section, but in practice it will

likelycorruptthe presence in the currentgeneration of concerns for the

distant future.

Third,a similar undermining occurs due to the presence within the

present generation of localized varients of the intergenerationalprob-lem. Forexample, governments and businesses aretypicallyheaded byelites whose time-horizons areextremelylimited.Theyhave a strongin-

centive to ignore altogether,or at least defer action on, problems whose

solutions demand high costs to be instituted on the present set of voters

and other politicallyinfluential groups for the sake of benefits to thosewho do not currentlyhave any political power.

Ifthese argumentsarecorrect,then there arestrong reasons for treat-

ing the intragenerational global warming problem as if it were a pris-oner's dilemma. This news is bad enough. But matters are made worse

by the fact that the usual solutions to prisoner'sdilemma situations also

tend to be undermined by the backgroundpresence of the intergenera-tional problem. First, the appeal to broad self-interest is problematic.

Althoughthe currentgenerationdo interactrepeatedly,so that there are

many gainsto be made

through cooperation,the

accessibilityof these

gains does not depend on cooperation about pollution. In fact, restrict-

ing pollution actually lessens the potential gains of cooperation, per-

haps even to the point where from a self-interested point of view it is

better to pollute and not cooperate than to cooperate and not pollute.So, the potential gains to be made from cooperation do not have the

same status as in normal prisoner'sdilemma cases.67Second, the prob-lem of monitoring compliance suggests that the appeal to fairness also

may not work.68Hardinmay be wrong that naturalselection eliminates

conscience, but it does seem plausible to say that, without an effective

method of ensuring compliance, the marketeliminates a sense of fair-ness, or at least a competitively disadvantageous sense of fairness. If

some producerscheat on energyconsumption, theywill have a compet-

67. Note that if the intergenerationalproblemis to be solved, the firstgenerationis al-

ready makingan absolute sacrificein terms of self-interest.Thismakes things difficult n

dealingwith noncooperators.Suppose they seem quite happyto pollute and not cooper-ate. Am I supposedto offerthem further ncentives to come back into the fold?But then I

am givingup bothmy potentialgainsfrompolluting(tothe futuregenerations),and some

of the furthergainsfromcooperation(tothe reluctantcooperators).Thisis a lot to ask.Just

overpollutingmyselfwill lookverytempting.

68. Theproblemis thatnoncompliance,even globalnoncompliance, is actuallya ben-efit to the present generationso the incentives to enforceare not high.

415




itive advantage, through exploiting the externalitycosts. Hence, overtime, they will eliminate more restrainedcompetitors.69

IX. IMPLICATIONS

Ifmy analysis of the inter- and intragenerationalproblems surrounding

global warming are correct, two of Hardin'smain claims are right afterall. First,strong coercive regulatory regimes may be needed in order to

stop overpollution,and so to address the pressing problem suggestedby

population growth.70 econd, the benign demographic transition hy-pothesis should be treated with suspicion.

Nevertheless, Hardin and I differ in our reasons for making these

claims, and this has practicalconsequences. In the first case, Hardinis

not correct about who the primary subjects of coercion should be. For it

is people in the richcountries who presentlycause most of the pollutionI have been concerned with, not those in the poor countries. Hence,while it is true that itwould be disastrous for thepoor countries to adoptthe more energy-and pollution-intensive lifestylesof theWest(and,as a

result there is reason toprevent

thishappening),

it is also true,and more

important,that even without theircontribution,the existingpatternsof

behavior in the West will have serious consequences, and must be

stopped. This should be the politicalpriority.7'In the second case, Hardin s skepticalabout the benign demographic

transitionhypothesis because he doubts the scientific evidence, and is

inclined towardsa strong evolutionaryaccount of human reproductivebehavior.ButIam skepticalbecause,althoughthe empiricalevidencefor

a decline in population seems compelling, I doubt that the expectedtransitionwill be benign. Forit comes at the priceof increased develop-

ment, and so increasedenergyconsumption and pollution.This locatesthe problemnot in the deep nature of human beingsand theirgermlines,but rather n ways of life which allof us could, and should, giveup.

69. Thecompetitivepressuresargumentdoes not work n allcontexts,but it does seem

operativehere. For some doubts about its general application,see Daniel HausmanandMichael McPherson, Economic Analysis and Moral Philosophy (Cambridge: Cambridge

University Press, 1996), chap. 4.

7o. In addition,note that the problemof overpollutionwould remain even in the ab-sence ofpopulationgrowth,since energyconsumptionper capitais on an upward piral.

71. Furthermore,t is probablya politicallynecessary prerequisite o preventing he de-

velopingcountriesfromfollowingaWesternpath.

416

gardner the real tragedy of the commons

Documents